This work is directed toward the development of improved statistical methods for the analysis of case-control studies. The adaptation of a proportional hazards disease incidence model to retrospective sampling is being studied. This work will permit the estimation of relative risks associated with multiple time-dependent exposure variables. Methods for odds ratio estimation based on logistic disease probabiliy models are also being adapted to case-control sampling. This work has shown that unmatched case-control data can be analyzed as if it had been collected prospectively, at least as far as asymptotic inference on the odds ration is concerned.